RI 9638

REPORT OF INVESTIGATIONS/1997

Distinguishing Motor Starts From Short

Circuits Through Phase-Angle
Measurements

UNITED STATES DEPARTMENT OF THE INTERIOR

UNITED STATES BUREAU OF MINES

U.S. Department of the Interior

Mission Statement
As the Nations principal conservation agency, the Department of the
Interior has responsibility for most of our nationally-owned public
lands and natural resources. This includes fostering sound use of our
land and water resources; protecting our fish, wildlife, and biological
diversity; preserving the environmental and cultural values of our
national parks and historical places; and providing for the enjoyment
of life through outdoor recreation. The Department assesses our
energy and mineral resources and works to ensure that their
development is in the best interests of all our people by encouraging
stewardship and citizen participation in their care. The Department
also has a major responsibility for American Indian reservation
communities and for people who live in island territories under U.S.
administration.

Report of Investigations 9638

Distinguishing Motor Starts From Short Circuits

Through Phase-Angle Measurements

By

Michael R. Yenchek, James C. Cawley, Jeffrey Shawn Peterson, and

Jeffrey L. Kohler

UNITED STATES DEPARTMENT OF THE INTERIOR

Bruce Babbitt, Secretary
BUREAU OF MINES
Rhea Lydia Graham, Director

International Standard Serial Number

ISSN 1066-5552

CONTENTS
Page
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analyses of high-voltage longwall power systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Approach to system analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Factors affecting short-circuit protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motor starts and short-circuit protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motor-start evaluations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Field tests of mine motors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data acquisition hardware and software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motor waveform analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motor-start phase-angle observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motor-start detection scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prototype circuit operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix.Circuit description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2
2
3
3
4
4
5
5
5
6
7
8
8
8
9
9
9
10

ILLUSTRATIONS

1.
2.
3.
4.
5.
6.
7.
8.
9.
A-1.

One-line diagram of a typical longwall power system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Coordination plot for 2,400-V longwall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coordination plot for 950-V longwall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measuring motor-start waveforms on a 950-V continuous mining machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Phase-to-phase voltage waveform for start of 950-V, 165-hp motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Line-current waveform for start of 950-V, 165-hp motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Superposition of motor-start voltage and current waveforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Current waveform distortion of pump motor on continuous mining machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Phase angle versus time during start of cutter motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prototype phase-angle detection circuitry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3
5
5
6
6
6
7
7
8
11

UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT

A

ampere

ms

millisecond

hp

horsepower

MVA

megavolt ampere

Hz

hertz

second

km

kilometer

volt

kV

kilovolt

percent

kVA

kilovolt ampere

degree

meter

Mention of any company name or product does not constitute endorsement by the National Institute for Occupational Safety and Health.
To receive other information about occupational safety and health problems, call 1-800-35-NIOSH (1-800-356-4674), or visit the
NIOSH Home Page on the World Wide Web at http://www.cdc.gov/niosh/homepage.html

DISTINGUISHING MOTOR STARTS FROM SHORT CIRCUITS

THROUGH PHASE-ANGLE MEASUREMENTS

Michael R. Yenchek,1 James C. Cawley,2 Jeffrey Shawn Peterson,1 and Jeffrey L. Kohler3

ABSTRACT
The Pittsburgh Research Center (PRC)4 investigated how the starting of induction motors may cause nuisance
tripping of short-circuit protection on coal mine power systems. This research had a threefold purpose: (1) to
determine the range of typical values for power system characteristics that affect short circuits and motor starts on
high-voltage longwalls, (2) to identify how motor-start waveforms differ from those for short circuits, and (3) to
devise a method to provide short-circuit protection without intentional time delays to account for motor starts.
Distribution voltage, transformer impedance, power center location, and motor size were found to critically influence
the magnitude of short-circuit and motor-start currents on high-voltage longwalls. An attribute of motor-start
signatures that distinguished them from short circuits was the relatively large phase angle between voltage and
current. Electronic circuitry was designed to detect phase angles and react to momentarily disable circuit breaker
action for motor starts. A prototype was successfully evaluated with an induction motor in the laboratory. By
minimizing intentional time delays in short-circuit protection, this technology will help ensure that surface
temperatures of energized electrical apparatus will not exceed gas or dust ignition thresholds when such faults occur.

Electrical engineer, Pittsburgh Research Center, National Institute for Occupational Safety and Health, Pittsburgh, PA.
Supervisory electrical engineer, Pittsburgh Research Center, National Institute for Occupational Safety and Health, Pittsburgh, PA.
3
Associate professor, Department of Mineral Engineering, The Pennsylvania State University, University, PA.
4
This research originated under the U.S. Bureau of Mines prior to transferring to the National Institute for Occupational Safety and Health in 1996.
2

INTRODUCTION
On August 27, 1992, the Mine Safety and Health Administration (MSHA) published proposed rules for high-voltage
electrical equipment used in longwall face areas of underground
5
coal mines [57 Fed. Reg. 39036 (1992)]. The intent of these
regulations is to reduce the likelihood of fire, explosion, and
shock hazards by citing requirements for electrical enclosures,
circuit protection, testing, and personnel protection.
The
operational limits for short-circuit protective devices are proposed
within section 75.814 of these rules.
Short-circuit protection for electrical apparatus in underground
coal mines is critical. When different phases of an electrical circuit
inadvertently come in contact, thousands of amperes may flow.
Currents of this magnitude can cause explosions and fires if
permitted to exist even for periods as brief as 1 s. The thermal
2
energy expended is directly related to I t [ANSI/IEEE standard
242, 1986]. Consequently, protective device settings must be
specified with a maximum sensitivity to current (I) and a minimum reaction time (t).
Unfortunately, extremely low current settings may interfere
with mining operations by inadvertently reacting to normal
transient events, such as the starting of motors. In response, the
proposed MSHA rules specify maximum time delays for shortcircuit protection of cables extending from the power center to
motor starters. In the future, the trend toward higher efficiency
motors
with
greater
peak
starting
currents
may

necessitate even longer delays.

However, in background
discussion of the rules, MSHA solicits comments regarding the
elimination of intentional time delays with a conjunctive increase
in current settings. This reflects the dilemma of short-circuit
protection, namely, that circuit protective devices should have
high sensitivity to faults, but not interfere with normal mining
operations.
Accordingly, MSHA requested a high-priority
research effort aimed at eliminating intentional time delays in
short-circuit protection.
The specific objectives of this research project were to
(1) determine the range of typical values for power system
characteristics that affect short-circuit and motor-start current
magnitudes on high-voltage longwalls, (2) identify how motorstart waveforms differ from those for short circuits, and (3) devise
a method to provide short-circuit protection without intentional
time delays to account for motor starts. Minimal reaction times to
short circuits will help ensure that the temperatures of energized
electrical apparatus will not exceed gas or dust ignition thresholds
when such faults occur.
Additionally, the capability of
distinguishing between motor-start and short-circuit events will
preclude nuisance protective device activation. Although the
initial thrust of the project focused on high-voltage longwalls, the
technology sub- sequently developed is applicable to low- and
medium-voltage mine power systems as well.

BACKGROUND
Protection against short circuits on mine power circuits is
typically provided by molded-case circuit breakers or vacuum
interrupters. The current magnitude thresholds or settings at which
6
these devices operate are specified either by regulation [30 CFR
75.601 (1992)] for low- and medium-voltage circuits or through
the MSHA approval process [Boring and Porter 1988] for highvoltage longwalls. The initial inrush currents, demanded by large
induction motors starting across the line, may exceed these settings
and activate the protection devices needlessly. Consequently, to
prevent nuisance tripping of the short-circuit protection, it is
desirable to seek a means to momentarily disable or inhibit the
protection device activation for a finite period following motor
energization.
5
6

Federal Register. See Fed. Reg. in references.

Code of Federal Regulations. See CFR in references.

A speed-sensing switch could logically be utilized to signal

motor start, but requires direct access to rotating parts, which is
sometimes not feasible or economical. In addition, this sensor may
be slow to operate where load inertia is high. This is also true for
induction-disk impedance or distance relays used primarily for fault
protection on transmission lines [Morley et al. 1982; Schulman
et al. 1978; DeCastro et al. 1995]. In the late 1970's, a system was
developed in the United Kingdom that monitored the phase angle
between voltage and current to distinguish between faults and
motor starts [Lord and Pearson 1980]. Despite promising results,
this technology has not been incorporated in short-circuit
protection for U.S. mines. Given the catastrophic potential of
inadequate electrical protection, it is imperative to reexamine this
problem to minimize any intentional time delays while maximizing
current sensitivity.

ANALYSES OF HIGH-VOLTAGE LONGWALL POWER SYSTEMS

To distinguish motor starts from short circuits, it was first
necessary to study the qualities of each event and identify the
factors that have critical impact on the resultant current. Be- cause
the impetus for this project resulted from proposed MSHA rules
involving high-voltage longwalls, the char- acteristics of these
power systems were evaluated at the outset. These analyses
enabled the construction of coordination plots that graphically
illustrated the problem of motor-start inter- ference with shortcircuit settings.
A database of coal mine high-voltage longwalls was obtained
from MSHA [Skorski and Checca 1993].
The electrical
information pertaining to 30 power systems was scanned to gain
familiarity with typical characteristics. A one-line diagram of a
typical longwall power system is shown in figure 1. Shearer motor
voltages were found to range from 950 to 4,160 V, with most
using 2,400 V. Shearer motor ratings varied from 370 to 1,480
hp, with the average at 920 hp. The cable supplying power to the
shearer motor typically was a 4/0 American wire gauge (AWG) that
ranged from 152 to 549 m long.
APPROACH TO SYSTEM ANALYSES
Commercial power system analysis software was used to model
these systems. In determining the factors that critically influence
short-circuit and motor-start currents on high-voltage longwalls,
several assumptions were made. Three-phase symmetrical and lineto-line faults were exclusively considered because ground faults are
sensed by separate devices. The three-phase symmetrical fault
closest to the circuit breaker represented the maximum fault current
that must be interrupted, whereas the line-to-line value, at the most
distant location from

the circuit breaker, represented the minimum value that must be

detected. A zero value of fault impedance was used to model
worst-case faults on high-voltage longwall power systems.
Under fault conditions, any motor connected to the faulted bus
will behave like a generator, supplying current to the fault for a
short time. As the voltage on the faulted bus collapses, the inertia
of the rotor and connected load will power the motor. Induction
motor contributions during faults were presumed to be equal to the
full-voltage starting current of the motor for up to two cycles
[Huenig 1982].
The effect of the mine power distribution system was the single
most difficult issue to resolve during this study, because the same
longwall system connected to two different dis- tribution systems
could perform very differently. The specific parameters of concern
were the short-circuit capacity at the main mine substation and the
impedance between the main substation and the longwall
transformer. The size of modern longwall power systems demands
that the power system be reasonably stiff (voltage regulation<10%),
and the size of main substations is such that short-circuit capacities
greater than 900 MVA are not uncommon in most of the United
States. Nonetheless, the impedance from the main substation to the
longwall transformer will be substantially larger as the mine
develops and up to 6 km of feeder cable is added.
Finally, one-third of the longwall power centers included in this
study utilized a three-winding transformer, with a high-voltage
secondary of 2,400 or 4,160 V and a medium-voltage tertiary. The
tertiary windings of longwall power center trans- formers were
neglected because their influence on short circuits was believed to
be negligible. In addition, future systems are likely to use the
tertiary winding only for ancillary equipment, such as pumps.

Figure 1.One-line diagram of a typical longwall power system.

FACTORS AFFECTING SHORT-CIRCUIT

PROTECTION

A primary goal of this study was to determine the range of

typical values for power system characteristics that impact short-

circuit protection settings of high-voltage longwall installations. It

was understood that transformer impedance and capacity, cable size
and length, and motor type and size would be important factors.
The variability of these parameters and their impact on short-circuit
protection settings were investigated.
The available short-circuit capacity at the main substation has a
major impact on the available short-circuit current at the longwall.
The stiff power systems common throughout most of the United
States can deliver 1,000 MVA or more. However, in some rural
areas, particularly in the Western United States, capacities are as
low as 100 MVA. Practically, this would more than halve the fault
current in a typical coal mine, e.g., 17,500 to 8,000 A. For this
study, a value of 1,000 MVA was used, not only because it is
typical of a large number of mines, but also because it represents
the more difficult case.
The mine distribution system to which the longwall power
system is connected ultimately limits the fault current at the
longwall. Any transformers between the utility connection and the
longwall power center add impedance that limits current flow, as
does the impedance of cables used to connect these points.
Initially, there may be as little as 600 m of cable between the main
substation and the longwall power center, whereas in later years
6 km may be in place before a new power borehole is required.
This additional cable impedance, along with the significant
transformer impedance, reduces available fault current by about
40%, e.g., 20,700 to 13,200 A at the secondary of the longwall
transformer.
The impedance of the longwall transformer is the single most
important variable within the longwall power system for
determining short-circuit currents. Impedances as low as 3% are
common, as are values up to 7%. The short-circuit current can be
25% greater for a 3% transformer impedance than for the same
system with a 7% transformer impedance. The transformer
capacity also affects fault current because larger transformers pass
more energy. However, given the typical range of transformer
sizes (3,000 to 5,000 kVA), this is a minor consideration compared
with the impedance [Morley et al. 1990].
The impedance of individual cables as determined by their
length, size, and type also influences the current magnitude. Type
SHD-GC cable is used universally in sizes of 2/0 and 4/0 AWG for
the armored face conveyor and shearer loads and often for the
crusher and stage loader. Face lengths range from 230 to 365 m;
however, the difference in total impedance caused by the additional
few hundred meters of cable is small. The variation in fault
currents, due to cable differences, is insignificant compared with
variations caused by transformer impedance.
The motor size and type affects the short-circuit settings
significantly. The motor's size, i.e., horsepower, and its design
parameters, especially its subtransient reactance, impacts the first
one to two cycles of fault current. As described earlier, the motor
contribution to the faulted bus can be significant and depends on
its subtransient reactance. Normally, this is only a concern when
specifying the close and latch rating of a circuit breaker or in
medium-voltage applications when using molded case breakers to
ensure that the breaker will be capable of opening under this
condition.

MOTOR STARTS AND SHORT-CIRCUIT

PROTECTION
Motor size and type also impact starting current magnitude.
Large, high-efficiency motors combine high breakdown torques
with significant locked rotor reactances. The result is that starting
currents for a 950-V motor may exceed 2,500 A, the maximum
allowable instantaneous circuit-breaker setting permitted under the
provisions of Title 30 without special authorization from the
Secretary of Labor [30 CFR 75.601 (1992)]. A scenario can be
envisioned where it becomes difficult to distinguish between
starting current and fault current on a particular bus. Each of the
following conditions will contribute to such a situation:
1. The transformer impedance is high, e.g., 7% rather than 3%.
2. The longwall transformer is located at its greatest distance
from the mine substation.
3. The distribution voltage is 7,200 V, rather than 13 kV or
higher.
4. The motor is large, of high-efficiency design, and/or has a
high breakdown torque.
Conditions 1, 2, and 3 will reduce the fault current, whereas
condition 4 will result in higher starting currents. Moreover, for
the sake of a worst case, it is assumed that conditions 1, 2, and 3
will not limit the starting current. In reality, however, they will
cause a significant voltage drop during starting, which will limit the
starting current (the motor is modeled as a constant impedance load
during starting and a constant power load during operation)
[Novak and Martin 1996].
A coordination plot for protecting the tail motor of the armored
face conveyor (AFC) will serve as an example. Figures 2 and 3
illustrate the 2,400-V and 950-V cases, respectively, in which a
500-hp AFC tail motor is connected to the motor control center via
a 305-m, 4/0 AWG cable. The short-circuit protection for the
2,400-V system consists of a very-inverse time overcurrent (TOC)
relay and a vacuum breaker, whereas an extremely-inverse TOC
relay and molded case circuit breaker are used on the 950-V circuit.
The maximum fault current results from a bolted three-phase fault
at the terminals of the breaker, whereas the minimum fault

Figure 2.Coordination plot for 2,400-V longwall.

current is a line-to-line fault at the motor bus. The motor

contribution is considered for both the minimum and maximum
values. The motor starting current is shown for a fully loaded
conveyor.
An analysis of figure 2 shows that it is easy to achieve
coordination, and even though a small time delay is introduced

Figure 3.Coordination plot for 950-V longwall.

to allow for starting, it is minimal. Moreover, any fault will cause

an instantaneous initiation of the circuit breaker. In figure 3, the
starting current due to the inrush component exceeds 2,500 A and
would require a special approval from MSHA. It can also be
observed that the margin between the starting current and the
minimum fault current is quite small.

MOTOR-START EVALUATIONS
The nature of how the starting of induction motors may cause
nuisance tripping of short-circuit protection has been analyzed.
Electrical safety can be enhanced if the circuit-protection device
recognizes and reacts only to short circuits. To devise a means to
discriminate between motor starts and short circuits, one must first
investigate how the waveforms of each event differ. Any
distinguishing characteristics may then be keyed upon through the
design of sensing circuitry. Accordingly, it was decided to record
the voltage and current waveforms of mine induction motors of
various voltages and evaluate their salient characteristics.
FIELD TESTS OF MINE MOTORS
To gain insight into motor-start signatures, field tests of a
variety of mine motors were conducted. These included motors
with application in both longwall and continuous mining, with
voltages ranging from 440 to 4,160 V and power ratings

from 10 to 450 hp. Because mine motors are typically started

across the line with full voltage, it was necessary to select test sites
with this capability. These included a rebuild shop with 440- and
550-V motors up to 150 hp. At another facility, a 4,160-V
conveyor motor was evaluated. Finally, the pump, conveyor, and
cutter motors of a 950-V continuous miner were analyzed
(figure 4). To ensure random motor energization relative to the
voltage cycle, 12 successive start recordings were made of each
motor. Recordings were usually made at the terminals of the
motor. However, recordings of the continuous miner motors were
made at both the motor terminals and in the load center that fed the
153-m trailing cable.
DATA ACQUISITION HARDWARE AND SOFTWARE
Transducers were used to record three-phase currents during
motor starts. The Hall-effect current sensors featured a frequency
response from direct current (dc) to 1,000 Hz to

Figure 4.Measuring motor-start waveforms on a 950-V con- tinuous mining

machine.

capture higher frequency transients that may be present. The

1,000-A split-core configurations had 5,000-V line-to-output
isolation. They were used in conjunction with matched signal
conditioners having 10-V outputs directly proportional to the
amplitude of the input signal. Tests with low-voltage motors
utilized 0 to 1,000-V voltage transducers with a proportional 0 to
10-V output. These sensors had a frequency range of dc to 5,000
Hz. At 1,000 V and above, custom potential trans- formers were
used to acquire the voltage waveform. The current signal
conditioners and voltage sensors were connected to groundisolation boards, 1,000-Hz programmable filters, and a 12-bit
analog-to-digital board. This board had the capability of scanning
up to eight separate channels and was equipped with simultaneous
sample-and-hold circuitry that eliminated phase shifts among
channels. The output of this board was digitally recorded in a
multichannel file using a 486-based laptop personal computer that
was made rugged for field use. A custom computer program was
written that allowed the user to choose the number of channels to
record at sampling rates up to 30,720 Hz per channel. During
acquisition, data were written directly to disk with data files saved
in binary format.
MOTOR WAVEFORM ANALYSES
The motor-start voltage and current data were subsequently
imported into a commercial data analysis package. Three-phase
waveforms of voltage and current were plotted versus time for each
of the motor tests. Typical plots are shown in figures 5 and 6. The
inrush current following circuit energization has the potential to
cause nuisance tripping of the short-circuit protection. It lasts for
a time period that depends on the motor characteristics and loading.
Closer examination of the time-varying signal revealed highfrequency oscillations within the

Figure 5.Phase-to-phase voltage waveform for start of 950-V, 165-hp motor.

Figure 6.Line-current waveform for start of 950-V, 165-hp motor.

first few milliseconds following energization. These pertur- bations

were later attributed to contact bounce. A fast Fourier transform of
the current waveforms displayed no frequencies other than the
fundamental and its harmonics. Consequently, the examination of
individual motor-start waveforms revealed no unusual
characteristics upon which a motor-start detection scheme could be
keyed.
Attention then turned to a study of the induction motor phase
angle during starting. The phase angle may be obtained by
superposing the voltage and current waveforms of a particular
phase on a common time scale (figure 7) and recognizing that
nominal 30 phase shifts were inherent with the phase-to-phase
connections of the voltage transducers. It should be noted that the
phase angles observed were not equivalent to those

associated with power factor. The angle in degrees between two

corresponding points of the waveforms may be derived by
considering that, at a frequency of 60 Hz, one 360 waveform
cycle is completed every 0.01667 s.
Typical inductive reactance-to-resistance (X/R) ratios for
induction motors, published in the Institute of Electrical and
Electronics Engineers "Red Book" [ANSI/IEEE standard 141,
1993], are in the range of 10 to 20. A fault on a circuit feeding an
induction motor essentially shunts that inductive load, decreasing
the phase angle between the voltage and current. Further, it is
known that the phase angle of an induction motor during start is
larger than that at full load [Schulman et al. 1978]. Accordingly,
the phase angle between the voltage and starting current must
theoretically be larger on any given mine motor circuit than that
resulting from a fault.
MOTOR-START PHASE-ANGLE OBSERVATIONS
Phase angles from start to full speed were derived from the
motor waveforms recorded during field tests. The voltage and
current waveforms were superposed and the time lag calculated at
the zero crossings from energization to the point where the motor
had attained full speed. Initial analyses involved nine unloaded,
Class H, 440- and 550-V induction motors at a rebuild facility.
The angle between voltage and current for the first and second zero
crossings was found to vary by as much as 50, not only among
each of the three phases for a given test, but also for any given
phase in successive tests. In addition, phase-angle variations were
observed even after the motor attained full speed. A number of
factors contributed to these phenomena. First, in many cases
immediately following energization, the current waveforms
exhibited a transient distortion that affected the exact time of the
zero crossing. High-frequency oscillations due to contact bounce
enhanced these irregularities. More importantly, the position of the
voltage wave of a particular phase relative to the time of
energization had an even greater impact on the phase angle. For
example, if the voltage of a given phase was approaching a zero

energization, the resultant current wave for that phase would

exhibit distortion during the first cycle (figures 5 and 6). Further,
some of the inconsistencies may be attributed simply to the fact that
all of the motors were tested on the shop floor without any
mechanical load on their shafts. Repeatable phase angles were
observed in subsequent tests of a loaded 4,160-V, 450-hp motor
at another facility and motors under load on a continuous miner.
Finally, as expected, there were differences observed in phase
angles during starts with motors of different horsepower or
manufacturers.
Next, the results of tests on continuous mining machine motors
were analyzed. During data acquisition, considerable waveform
distortion of the pump motor current was noted (figure 8). Again,
there was some variation among calculated phase angles during the
first two cycles, but not to the degree observed with unloaded
motors. Despite the waveform distortion, following the first two
cycles and continuing to running speed, phase-angle magnitude
behaved much more predictably over time among phases and
successive tests. Figure 9 shows that the phase angle from start to
full speed for the cutter motor declined almost linearly as measured
at the power center. Not surprisingly, because trailing cables
exhibit a lower X/R ratio compared with that of motors, the phase
angles calculated at the motor terminals during starts were on
average 5 greater than those obtained from tests at the power
center.
These analyses provided insight into the behavior of the phase
angle between the current and voltage of a motor during start.
They showed that, for a loaded motor after the first two cycles, the
phase angle was relatively predictable. Con- sequently, the phase
angle can be a reasonable basis for a methodology designed to
distinguish between motor starts and short circuits on mine power
systems.

Figure 8.Current waveform distortion of pump motor on continuous mining

machine.

Figure 7.Superposition of motor-start voltage and current waveforms.

crossing and about to change polarity at the moment of motor

Figure 9.Phase angle versus time during start of cutter motor.

MOTOR-START DETECTION SCHEME

The proposed MSHA rules for high-voltage longwalls permit
a maximum time delay in short-circuit protection of 0.25 s or 15
cycles [57 Fed. Reg. 39036 (1992)]. To minimize electrical fire
hazards, a motor-start detection scheme should react much faster,
ideally with no intentional time delay. Initially, it was believed that
commercial phase-angle meters might be suited for incorporation
into a detection system. However, the typical response times
specified for these instruments was found to be 100 ms, which is
too slow for consideration. Minimal time delay is best achieved
through a digital-based design that does not rely on computational
processes or integration of 60-Hz-power waveforms. Accordingly,
the approach was to devise a means to detect the time between each
zero crossing of the voltage and current signals beginning with the
first half-cycle.
PROTOTYPE CIRCUIT OPERATION
The prototype circuit is described in detail in the appendix. The
circuit can monitor the current and voltage waveforms for one
phase of a three-phase circuit. A suitable potential transformer is
required for a phase-to-phase voltage measure- ment. To accurately
measure electrical degrees during each half-cycle, a crystalcontrolled oscillator was selected as the time reference. Clock
pulses generated during zero crossings of voltage and current are
counted and compared with a preset (phase-angle) delay. If the
angle between voltage and current exceeds the setting of the circuit
and current magnitude is high, the device sends a signal to inhibit
circuit breaker action.
The detector circuit was tested using a 50-hp, 460-V induction
motor in the laboratory. The motor had an unloaded generator
coupled to its shaft. The detector's activation setting was adjusted
to within 5 of the phase-angle characteristics recorded during prior
motor starts. Once set, the detector contacts were observed to drop
out during subsequent motor starts. Occasionally, they remained

closed momentarily at the outset of a test, but only for a period

corresponding to the first two cycles of motor-start current. During
this initial period, switching transients may be expected.
APPLICATION
The prototype detector can have utility on three-phase power
systems where the starting of induction motors causes nuisance
tripping of the short-circuit protection. In lieu of incorporating an
intentional time delay or increasing the magnetic trip setting of the
circuit breaker, the detector may be used to momentarily inhibit
circuit breaker action during motor starts. It is en- visioned that the
detector would consist of three identical circuits, one for each
phase. Located near the circuit breaker, the device would
temporarily inhibit breaker action only if low phase angles and
high currents were indicated for all phases. This would preclude
disabling the circuit protection for phase shifts associated with lineto-line faults. Inherent shifts of 30 due to phase-to-phase
connections of the voltage transducers would simply be added to
the motor-phase shifts when setting the detector's phase-angle
activation point. It is recommended that the activation point of the
current magnitude detector be adjusted to the required magnetic
setting of the circuit breaker. Consequently, the device would
remain inactive for currents less than the circuit breaker setting.
When phase currents exceed this level, a built-in two-cycle delay
in detector activation (not part of the original design) would
prevent false indications for the initial period of motor starts.
Because the phase-angle characteristics of motors depend on motor
voltage, class, manufacturer, and horsepower, presetting of the
detector would be impractical. In application, the device activation
point plus some margin would be determined by starting the
protected

motor a dozen times. This would account for any transients

resulting from the randomness of the instant of energization relative
to the voltage cycle.

SUMMARY AND CONCLUSIONS

The Pittsburgh Research Center conducted an investigation to
improve electrical safety in coal mines through research into shortcircuit protection of mine power systems. Distribution voltage,
transformer impedance, power center location, and motor size were
found to critically influence the magnitude of short circuits and
motor starts on high-voltage longwalls. The need for accurate and
rapid fault detection was confirmed through computer models of
power systems. Voltage and current waveforms of mine induction
motors were recorded during field tests, and their salient
c h a r a c t e r i s t i c s
w e r e
evaluated.
An attribute of motor-start signatures that distinguished them from short circuits was the relatively large phase
angle between voltage and current. This was the focus of
subsequent efforts to devise a means to discriminate between motor
starts and short circuits. Electronic circuitry was designed to
differentiate between phase angles associated with

motor starts and faults and react to momentarily disable circuit

breaker action for motor starts. A prototype was successfully
evaluated with an induction motor in the laboratory.
By devising a method to distinguish between motor starts and
short circuits, any intentional time delays in response to short
circuits can be virtually eliminated. Rapid response to short circuits
will help ensure that the surface temperatures of energized electrical
apparatus will not exceed gas- or dust- ignition thresholds when
such
faults
occur.
In
addition,
the
capability of differentiating motor starts from short circuits will
preclude nuisance protective device activation. Ultimately, this
work may lead to revisions to high-voltage longwall approval
guidelines and improvements in circuit breaker designs. However,
the technology should be applicable to motor circuits at all voltage
levels.

ACKNOWLEDGMENT
The authors acknowledge the assistance of Elio L. Checca of the
MSHA Technical Support Group, Pittsburgh, PA, who

provided the results of electrical surveys of high-voltage longwalls

for use in this project.

REFERENCES
ANSI/IEEE standard 141 [1993]. Recommended practice for electric power
distribution for industrial plants (IEEE red book). Institute of Electrical and
Electronics Engineers, Inc.: New York, NY.
ANSI/IEEE standard 242 [1986]. IEEE recommended practice for protection
and coordination of industrial and commercial power systems (IEEE buff book).
Institute of Electrical and Electronics Engineers, Inc.: New York, NY, p. 344.
Boring CM, Porter KJ [1988]. Criteria for approval of mining equipment
incorporating on-board switching of high-voltage circuits. In: Proceedings of the
Ninth WVU Mining Electrotechnology Conference. Morgantown, WV: West
Virginia University, pp. 267-274.
CFR. Code of Federal regulations. Washington, DC: U.S. Government
Printing Office, Office of the Federal Register.
DeCastro JE, Beck RT, Cai C, Yu L [1995]. Stall protection of large
induction motors. In: IEEE Transactions on Industry Applications. Vol. 31,
No. 5, pp. 1159-1166.
57 Fed. Reg. 39036 [1992]. Mine Safety and Health Administration:
proposed rules.
Huenig WC [1982]. Calculating short-circuit currents with contributions from
induction motors. In: IEEE Transactions on Industry Applications. Vol. IA-18,
No. 2, pp. 85-92.

Lord H, Pearson FK [1980]. Ultra-sensitive instantaneous fault protection for

cage induction motor supplies. Mining Technology 62:219-223.
Morley LA, Tomlinson JN, Moore G, Kline DE [1982]. Portable trailing
cables, splices, and couplers design and installation considerations. University
Park, PA: The Pennsylvania State University. U.S. Bureau of Mines contract
No. J0199106.
Morley LA, Novak T, Davidson I [1990]. The application of 2,400 V to
longwall face equipment. In: IEEE Transactions on Industry Applications.
Vol. 26, No. 5, pp. 886-892.
Novak T, Martin JK [1996]. The applications of 4,160 V to longwall face
equipment. In: IEEE Transactions on Industry Applications. Vol. 32, No. 2, pp.
471-479.
Schulman JM, Elmore WA, Bailey KD [1978]. Motor starting protection by
impedance sensing. In: IEEE Transactions on Power Apparatus and Systems.
Vol. PAS-97, No. 5, pp. 1689-1695.
Skorski DF, Checca EL [1993]. High-voltage longwall database. In:
Conference Record of the IEEE Industry Applications Society (IAS), 28th IAS
Annual Meeting. Toronto, Ontario, Canada: pp. 2436-2441.

10

APPENDIX.CIRCUIT DESCRIPTION
The circuit described here (figure A-1) can monitor the current
and voltage waveforms for one phase of a three-phase circuit. A
suitable potential transformer is required for a phase-to-phase
voltage measurement.
Referring to figure A-1, the 120-V
secondary voltage is fed to pin 5 of IC1 through R11. IC1
develops an output pulse on pin 4 each time the input voltage goes
through zero. Similarly, a current transducer supplies a 120-V
signal that corresponds to the magnitude of the line current to pin
5 of IC7 through R12. Line current magnitude is also measured by
IC's 11 and 12, which are described later. Pin 4 of IC12 develops
an output pulse during each current zero crossing.
To accurately measure electrical degrees during each half-cycle,
a crystal-controlled complementary-metallic-oxide-semiconductor
oscillator, producing clock pulses at 32,768 Hz, was selected as a
time reference. There are 21,600 (60 Hz 360 electrical
degrees/cycle) electrical degrees per second. In a full 60-Hz cycle,
there are 546 (32,768 clock cycles/second (1/60)
second/electrical cycle) clock cycles. By proportion, the number
of clock cycles corresponding to a desired time delay (or phase
angle) between voltage and current waveforms is shown below:
Electrical
degrees . . . .

150 120

70

60

45

30

20

10

Clock
cycles . . . . . .

227 182 137 122 107

91

69

46

31

16

90

80

The voltage zero-crossing pulse from IC1 pin 4 sets the set-reset
flip-flop, IC2. Once set, IC2 enables IC3 to gate clock pulses from
IC10 into the binary counter, IC4. IC4 counts the clock pulses
until a current zero-crossing pulse is received from IC7 pin 4.
When the current zero-crossing pulse is generated, it resets the setreset flip-flop, IC2, and resets the counter, IC4, to zero. In
addition, when IC2 is reset, the flow of clock pulses to the counter,
IC4, is stopped until a voltage zero-crossing pulse begins the
process over again during the next half-cycle.

Switch SW1 is connected to the output of counter IC4. It

contains eight on-off switches that can be configured to indicate
when counter IC4 has reached a preselected number of clock pulses
during any half-cycle. By placing any switch in the "on" position,
the counter digit is constrained high. By using all eight switches
in this manner, an eight-bit binary number can be formed that is the
set point for the clock pulses that arrive between the voltage zerocrossing (set) pulse and the current zero-crossing (reset) pulse.
For example, to detect a delay of 45 electrical degrees, 69 clock
pulses must be counted. With the switches set to 01000101 (69 in
binary), IC8 delivers a pulse when the counter output equals
01000101. If the current zero-crossing (reset) pulse arrives before
count 01000101, IC8 will not deliver an output pulse during that
half-cycle, indicating a small (resistive) phase angle. Thus, this
device measures whether some preset phase angle is equaled or
exceeded during each half-cycle.
When the counter output equals the set point, the combinational
logic of IC's 5, 6, and 8 causes a high pulse to be sent to the
retriggerable one-shot, IC9. IC9 produces a 9-ms output pulse that
holds transistor Q3 and relay RL1 latched. If the voltage zerocrossing (set) pulse and the current zero-crossing (reset) pulse arrive
too closely together (small phase angle), IC4 cannot reach its set
point. A retriggering pulse is not delivered to IC9 during that halfcycle, causing it to switch to a low output and Q3 and RL1 to drop
out. When RL1s contacts drop out, continuity is lost in the relay
holding in the main circuit protective device, causing it to drop out
also.
In addition to detecting phase angle, IC's 11 and 12 measure
current magnitude. Half of IC11 scales the current waveform
arriving from the current transducer. This scaled waveform is fullwave rectified by IC12 and compared against an adjustable set
point in the other half of IC11. Logically combined with the
output of IC9, the prototype allows the circuit breaker to activate
on a combination of both high current and small phase angle.

Figure A-1.Prototype phase-angle detection circuitry.

11

12

Figure 8.Current waveform distortion of pump motor on continuous mining

machine.